Project
| # | Title | Team Members | TA | Documents | Sponsor |
|---|---|---|---|---|---|
| 15 | Smart Scooter Battery Management System |
Namin Shah Nicolas Nkiere Raj Lulla |
uma Lath | design_document1.pdf design_document2.pdf design_document3.pdf final_paper1.pdf photo1.jpg photo2.jpg presentation1.pptx proposal1.pdf video1.mov |
|
| # Smart Scooter Battery Management System Team Members: - Namin Shah (namins2) - Raj Lulla (rlulla2) - We are still looking for a third teammate # Background This idea is the foundation of our student run startup company Haylon Technologies. We are currently a member of the College New Venture Challenge and are developing a new type of ‘smart’ battery management system capable of taking two batteries of different chemistries and making them work in conjunction with each other. On a broad level, the goal is to create a ‘hybrid’ battery that can utilize the strengths of one type of battery chemistry to alleviate the weaknesses of the other. Combining a battery chemistry with a long cyclic stability and a high specific power, with a complementary cell that has a lower cyclic stability but a higher specific energy can enable a system that has the strengths of both chemistry types. # Problem Rechargeable batteries are inherently limited due to properties specific to their chemistry and production methods. The most common of which are traditional lithium ion batteries. Battery chemistries such as Lithium-Titanate can charge and discharge much quicker than the dominant lithium-ion batteries but have lower energy densities, whereas battery chemistries such lithium sulfur exhibit much better energy density than standard lithium ion yet struggle with the problem of cyclic stability. Existing solutions fail to bridge the gap between high energy cells and high power cells. The reason lies in the fact that high energy cells traditionally struggle with cyclic stability, meaning that they do not retain their capacity after many charge/discharge cycles. This happens due to a multitude of reasons. Whilst some of those reasons are inherent to the battery, others such as the discharge curve that battery experiences during each cycle and the depth of discharge can be controlled and even optimized by a smart battery management system that can selectively shift power draw to another cell. Simply stacking a high energy cell and a high power cell in parallel (assuming oring diodes are used and cell voltages are the same), would not ensure that the high energy cell follows any particular discharge curve. The first being that the high energy and high power cells do not have internal resistances that are at all proportional to their relative cyclic stability, and their maximum discharge rate. Put simply all of these batteries have similar internal resistances, unlike capacitors which have minimal internal resistance relative to batteries. This is why output capacitors are effective in such a simple arrangement, as buffers to deal with large current spikes.The problem with capacitors in general as energy storage elements, is their incredibly poor energy density which makes them impractical for usage in consumer electronics as a significant energy storage system. Rechargeable batteries, which universally contain energy densities several orders of magnitude above most capacitors, are really the only viable options in practice. These current spikes are what cause a battery's discharge curve to have a chaotic shape. Say in an electric scooter, during a period of acceleration where the cells current draw spikes, although the high power cell would absorb some of this spike, the high energy cell would still experience a larger current draw during this period. Whilst the impact that this has is minimal in the scope of one cycle, the constant fluctuation in current draw would tremendously impact the total lifespan of the batteries. So in order for two rechargeable batteries with different chemistries to work in conjunction a more sophisticated solution is needed. A solution where the high power cell (which can survive for several times as many charge cycles) can be intelligently be used as a power buffer to control the discharge curve of the high energy cell. # Solution The solution is a ‘smart’ battery management system as mentioned earlier. Our goal is to construct a battery that consists of two separate energy buffers: one which consists of a high energy battery chemistry and the other which consists of a high power battery chemistry. A low power microcontroller will manage the high side switching circuits which connect each buffer to the load, and also control the duty cycle of a specially designed power converter which brings charge from the high energy buffer to the high power buffer. The microcontroller should be able to precisely control the timing and more importantly the speed at which the high energy cell charges the high power cell. This is important because we intend to develop a fairly sophisticated control algorithm to manage the discharge curve of the high energy cell during a real life use case in an actual electronic device. In order to do so, our smart bms system will measure current draw, and the state of charge of each buffer at all times and evaluate the exact timings for each buffer to supply power to the load whilst simultaneously determining how quickly the high power buffer recharges. To develop this algorithm we intend to build the management system and report these three pieces of data constantly via bluetooth serial or wifi. On the backend with this data we can create the ‘true’ discharge curve of each cell during each cycle. With that knowledge we can begin to train an algorithm to best emulate an ideal discharge curve. The ultimate goal is to get the high energy cell to most closely follow what would be a constant discharge over time as opposed to a constantly changing curve that normally exists. To demonstrate this we intend to build a smart battery pack for an electric scooter. We chose this use case as the battery size is big enough for the power usage of our digital and analog electronics to not significantly matter, yet not too big as to exceed what would be practical for a senior design project. Also an electric scooter requires large current spikes quite often and during regular use, is not pushed to its capacity limit with each and every cycle. This creates the perfect system to demonstrate how our smart bms would in practice extend the usable lifespan of high energy rechargeable batteries, as it would detect these conditions and adjust the power buffers to best ensure long term battery health. As is such, the smart bms would also contain a gyroscope and accelerometer to help create the switching algorithm The long term goal is to create the technology that would enable a high energy battery chemistry such as lithium sulfur which currently struggles with cyclic stability, to be usable. However for this project we simply want to demonstrate that a hybrid battery system that could solve this problem, is possible. So for our particular scooter battery pack, we intend to use readily available Lithium Nickel Cobalt Aluminum Oxide cells (Li-NCA) as our high energy buffer and lithium iron phosphate (LFP) cells for our high power buffer. Both are commercially available on Digikey and other common electronic vendors and exhibit complementary strengths and weaknesses. Even though we are utilizing off the shelf batteries to simply demonstrate this idea, we still intend to create a scooter battery pack that matches the maximum discharge rate of any regular LFP battery system, whilst having a higher total energy density by virtue of the high energy Li-NCA battery. # Solution Components ## Subsystem 1 - Switching and inter battery charging circuit This circuit would enable an embedded-system to drive low resistance FET based high side switches for each sub-chemistry battery. As mentioned above, it would be able to alternate which chemistry battery is in use at any given moment. We intend to use a low power STM based microcontroller to drive this switching, based off the input data it collects from various sensors such as current sensors, voltage monitors, and accelerometers. Along with the switches, this microcontroller will drive a power converter that enables the high energy battery to deliver charge to the high power element, in a controllable manner. ## Subsystem 2 - Hybridization Algorithm The hybridization subsystem consists of three main components: 1. Collecting data: we need to collect as much data as possible on battery usage in our use case, in this case an electric scooter. 2. Analyze the data: we need to analyze the patterns in this data to collect information such as how often power spikes occur, how long power spikes last for, how long constant power draw lasts for, and charging tendencies. 3. Develop the algorithm: using this analyzed data, we need to develop the algorithm that controls our hybrid battery. We need the algorithm to intelligently manage charge between the two batteries so it is prepared for a power surge whenever one occurs while maintaining optimal charging and discharging curves for the high energy element. # Criterion For Success We realize that improving the usable lifespan of high energy batteries is somewhat arbitrary and would be hard to quantify at the end of the semester, so below are some measurable criteria of success that we will aim for. - Successful high side switching circuits that allow the microcontroller to alternate between energy buffers that supply to the load at will. - A custom charge controller that can charge the high power cell via the high energy cell during operation. - The microcontroller should be able to precisely control the timing and more importantly the speed at which the high energy cell charges the high power cell. This should be a feature of the aforementioned charge controller. - A detailed analysis of experimental data points including but not limited too: Current draw, state of charge, and speed/direction of the scooter to create an advanced switching algorithm to shape the discharge curve of the high energy cell. - A successful version of the switching algorithm that visually improves the shape of the discharge curve of our Li-NCA cell, when compared to a control (which will be obtained by simply letting just the Li-NCA drive the scooter across a regular charge cycle) |
|||||